SAFC Hitech moves phase change memory closer to commercial use

The company reported that extensive development work has been conducted with both the precursors and with the use of conventional metal-organic chemical vapor deposition (MOCVD) techniques to deposit them, resulting in the successful deposition of device-quality GST, which are believed to represent a major step towards achieving a commercially-viable solution to address the aggressive memory device scaling issues faced by the semiconductor industry to keep pace with Moore's Law.

By Ann Steffora Mutschler, Senior Editor -- Electronic News, 9/4/2008

For use in high volume manufacturing phase change memory (PCM) applications, St. Louis, Mo-based semiconductor manufacturing materials company SAFC Hitech said today that it has made significant progress in developing germanium antimony telluride (Ge_xSb_yTe_z or GST) precursors.

The company reported that extensive development work has been conducted with both the precursors and with the use of conventional metal-organic chemical vapor deposition (MOCVD) techniques to deposit them, resulting in the successful deposition of device-quality GST, which are believed to represent a major step towards achieving a commercially-viable solution to address the aggressive memory device scaling issues faced by the semiconductor industry to keep pace with Moore’s Law.

PCM is a non-volatile computer memory that leverages the unique behavioral properties of chalcogenide compounds to allow scaling of feature size further than is possible with conventional flash memories, which translates to greater storage capacity and superior performance for memory devices.

Chalcogenide compounds, such as GST, are very attractive materials for PCM and have already been used as the basis for optical storage media and prototype PCM devices, SAFC Hitech said.

“Until now, PCM materials have generally been deposited by sputtering or other physical vapor deposition (PVD) techniques, which are line of sight methods and have inherent weaknesses relating to uniformity of deposition,” explained SAFC Hitech CTO Ravi Kanjolia, in a statement.

“Vapor phase deposition techniques, such as MOCVD, offer several advantages in relation to GST precursors, in particular, a better step coverage for deposition on patterned substrates, industrial scaling and high compositional control. Furthermore, we have achieved advances in precursor chemistries that allow similar layers to be deposited using conventional MOCVD, without the need for an activation process,” he added.

SAFC Hitech also said it has been a participant in the European Commission-supported CHEMAPH project since 2005, which is a consortium set up to examine deposition methods for GST films. Also, researchers at the company’s Bromborough, UK-based facility have been investigating a variety of GST sources suitable for MOCVD, and have matched the physical properties of each metal precursor to enhance efficiencies at the desired growth parameters.

Kanjolia continued, “Variations in cracking efficiencies were one major hurdle that we had to overcome. After extensive work to synthesise a number of different chemicals and characterize their physical properties, a combination of sources was found with a much improved match of thermal stability to allow decomposition to the same degree when simultaneously introduced to the deposition reactor chamber.”

The chemicals of choice were found to be Ge(NMe₂)₄, Sb(NMe₂)₃ and iPr₂Te and with these identified, SAFC Hitech said it then developed synthesis protocols to allow the isolation of high purity product in both small and large laboratory scale equipment.

These materials are now available to customers with guaranteed quality by state-of-the-art in-house analysis with samples shipped to various centers, and collaborations with partners taking place to test the different combinations. Recent growth trials have resulted in successful deposition of device-quality GST using nitrogen as a carrier gas, the company pointed out.

“While a full process to make MOCVD devices remains to be demonstrated on anything other than very small research structures, the quality of the films on flat substrates is improving, and the precursor chemistry is ideally-suited. The next challenge is to get the correct parameters in place to control the growth and lay down the correct layers in the correct structure. The temperature window with this process remains critical, and highlights both the difficulties associated with this system and the need for advanced precursors to move forward with integration into future phase change memory applications. We are confident that our sources will allow the development of next-generation devices to maintain the speed of performance enhancement required to meet market targets,” Kanjolia concluded.